Digital Twins in Paint Manufacturing: Data Engineering Approaches for Process Optimization

Abstract

Digital Twins in Paint Manufacturing: Data Engineering Approaches for Process Optimization

Digital twins are virtual replicas of a physical entity. A digital twin can mimic the properties of its physical counterpart. Factors such as data fidelity, availability, and the required refresh rates are crucial considerations that provide the mapping between the digital and physical entities. In this chapter, we explore the concept of digital twins, their underlying philosophies, implementation strategies around data engineering foundations relevant to paint manufacturing, as well as their implications for process optimization. The importance of engineering the flow of existing and novel data sources to populate digital twins that enhance optimization techniques is a consistent theme throughout the chapter. We use paint manufacturing as a representative application area throughout the chapter. Paints are multicomponent fluid systems whose rheology is a predictor of many important properties such as printability, scrub resistance, and anti-corrosion properties. We argue that digital twins are a key enabler that allows the incorporation of closed-loop optimization techniques to ensure that there is always a mapped control action based on physical measurements, that signals the appropriate actuators to drive the paint-making process towards its optimal setpoints. Such control actions direct the system bots towards achieving the realized optimal paint solution when there is a client toleranced deviation in paint property, based on process flow recommendations made by the digital twin. The chapter describes how data ledging of the past data generated from the lab, pilot, and production equipment as well as the recommendation-based deployment of the digital twins can augment and work seamlessly with both existing empirical models as well as mechanistic understanding of the paint-making process leading to a significant gain in property realizations. We also discuss the potential transformation in process optimization in how companies operate, as they move from traditional thermal systems of risk assessment for business decisions towards a digital economy where there is constant vigilance of optimization in every stage of the asset lifecycle.